Apolipoprotein E epsilon4 (APOE-?4) allele carriers are known to be at substantially greater risk for cognitive decline and Alzheimer's disease (AD). Yet, APOE-?4 allele inheritance is an imperfect predictor of who will develop clinical symptoms of the disease, suggesting that modifiable lifestyle factors such as exercise may moderate its influence on disease progression. Our team is uniquely qualified, and we have published several preliminary studies showing that physical activity may offer protection for APOE-?4 allele carriers from AD- related neurodegeneration and cognitive decline. Interventions, such as exercise, that even modestly delay the onset of cognitive impairment or improve cognitive function in healthy APOE-?4 carriers will have a major public health impact. It is not yet known, however, if exercise prospectively modifies the disease trajectory in healthy asymptomatic older adults who are at increased genetic risk for AD. The focus and innovative aspect of our proposal is to test the hypothesis that exercise training will improve the efficiency of neural networks during memory retrieval, increase resting cerebral blood flow, neural network connectivity, and cortical thickness, and improve episodic memory performance in APOE-?4 allele carriers. There are three key knowledge gaps regarding exercise as a primary prevention of cognitive decline in those at genetic risk for AD. First, it has not yet been firmly established that exercise improves the function and efficiency of neuronal networks during cognition, including memory retrieval, in APOE-?4 allele carriers. Second, it is unknown if the neurotrophic and increased resting cerebral blood flow effects of exercise extend to APOE-?4 allele carriers. Third, it has not been demonstrated that an exercise intervention will have lasting effects that delay cognitive decline or conversion to MCI. The novel and distinguishing feature of our proposal is to address the first two knowledge gaps with MRI and cognitive outcomes after exercise training in cognitively intact older APOE-?4 allele carriers. This pilot clinical trial will inform a future trial addressing the critical question on the long-term effects of exercise in ?4 carriers. Cognitively intact APOE-?4 allele carriers will be randomly assigned to 6-months of either supervised moderate intensity aerobic exercise training (ET) or supervised flexibility exercise control (FC). The ET and FC each contain a group based exercise component and are run in retirement communities. Our primary aims are to compare pre-intervention to post-intervention changes in 1) MRI biomarkers; and 2) episodic memory performance measured by the Rey Auditory Verbal Learning Test (RAVLT). We hypothesize that after ET compared to FC, brain activation during memory retrieval will be reduced, resting cerebral blood flow and functional connectivity will increase in frontal regions, and episodic memory performance will improve. Outcomes in response to the intervention will be measured at baseline and 6 months. Our famous name discrimination task has several advantages to track the effects of any intervention on neural network efficiency and activates brain regions associated with the ?default mode network?, which is known to harbor amyloid and to be disrupted with progression to AD.